User talk:Vanished user adhmfdfmykrdyr/Blocks on English Wikinews

From Wikinews, the free news source you can write!
Jump to navigation Jump to search

Actionable thoughts[edit]

Assuming this is the potential basis for future community action, a couple of things strike me as actions in response:

  • Possibly removing 3RR as a blockable cause. Article archiving and the review process mean this type of block is unlikely to happen. The last time it happened was 2008.
  • Inappropriate user names are often tied to socking. Socking also often appears to be tied up to advertising and spam. A rationale could be written explaining that better connects these and puts them into a similar blocking category. (Though the underlying behaviors/motivations between the first group and the second group are the second, even if the outcomes are similar.)
  • Consider consolidating block rationales into two large groups : "block per community behavior" or "block per content". Then leave things like open proxy as a different often used, non-overlapping block rationales.

--LauraHale (talk) 00:26, 6 May 2013 (UTC)

Response on research list.[edit]

This is a copy / paste response of mine that I posted on the research list in response to another person's critique of this analysis. Posting it here because I think it explains some things. I may also want to reference it in case I decide to tweak the research. --LauraHale (talk) 00:36, 6 May 2013 (UTC)

I wanted to respond, but before I did, I wanted to consult with another two administrators on English Wikinews to make sure I have the background correct regarding the project's history, adoption of project policies and if the conclusions I drew based on the quantitative research I did on our blocking history matched with their perceptions.

The first issue with your critique of my research is it appears to be entirely premised on the uncritical acceptance of English Wikipedia's "Assume Good Faith" (AGF) as a default world view of how things were, are and should be without bothering to delve into how AGF can and does function in terms of community behavior. Indeed, this extends beyond your critique of English Wikinews's blocking history but is possibly a fundamental flaw in your own research because you fail to acknowledge how AGF may be playing an underlying role in English Wikipedia's community development as it relates to changes in blocking patterns over time. It might be reasonable to conclude that the increase in indefinite blocks on English Wikipedia that are not overturned is not evidence of a spam problem, but evidence of a community shift away from the AGF policy as fundamentally unworkable. (It actually begs the question of if and what are the changes in overall block length on English Wikipedia, when did they occur, and what other things were taking place in the community. If your research is contextualized against other projects which could show no increase in spam problem, it would support a conclusion regarding community changes in understanding of AGF.)

This is a key difference when looking at English Wikinews is to understand there are distinct differences in how the local culture functions because of "Never Assume" (NA). English Wikipedia makes the assumption that based on AGF that a person who inserts copyright violations or wishes death on a contributor or inserts libelous material is acting in good faith and does not understand policy. As was pointed out to me, NA is not an official policy but a governing thought underlining our community. Thus, we do not make assumptions based on person's motives but instead look at their actions. If you insert a copyright violation, wish death on another contributor or insert libelous material, these actions are fundamentally harmful to the community. We cannot assume you meant well. We can take steps to avoid these and protect the community from harm. (Copyright violations, destructive community members not participating with the shared goal of working on news, actively lying and opening the project up to legal problems.) Thus, our lack of willingness to tolerate destructive behavior means we have a much more consistent pattern of indefinite blocking compared to English Wikipedia. (Indeed, this can clearly be seen because the number of overturned blocks fell after the community stabilized and NA was adopted. It would be an interesting question to see the rate of overturned indefinute blocks on English Wikipedia before and after the adoption of AGF.)

Beyond that, your critique suggests that decisions to change page protection are not community policies and user activity. This appears to be pretty absurd on some level. If English Wikipedia decided that after an article was assessed as Good Article, it would no longer be open to editing by any contributor but an administrator, what do you think the impact on the community and user activity would be? Page protection does impact community and how users behave. The review process also is a community function and user activity. Again, going back to English Wikipedia, hypothesize the impact on how the community and its users would function if the community decided that articles would only go live on English Wikipedia if they could pass a Good Article assessment, with the community having determined the requirement for requiring such a policy was to insure compliance with verifiability, copyright, neutrality and English Wikipedia's style guide. I would be incredibly shocked if that happened on English Wikipedia if there was not a fundamental shift in how blocks were implemented. Thus, these underlying community functions for page protection and article publishing do play a role.

The overall point of your critique also appears to be that the research would have been better served by doing a month by month analysis. Ours is not sample like yours, but instead includes the full block log. There is a question of diminishing returns of doing a month to month when your overall block log is only 15,100 or so, with roughly 6,800 of those blocks made in one month by one administrator that specifically dealt with open proxes. Once those are taken out of the equation, we have roughly 8,300 blocks which is around 100 months. On average, that would mean 83 blocks a month. When you start breaking these blocks down further by category, it begins to look increasingly less meaningless. 14 copyright related blocks spread over 100 months instead of 8 years and 4 months offers little value analysis wise. It does not allow for an examination in major events that impacted community because of the relative size of the sample. --LauraHale (talk) 00:36, 6 May 2013 (UTC)

Was this posted on reply to me? I'd note that you don't seem to have ever actually hit send (indeed, I was surprised not to get a response.) Ironholds (talk) 12:09, 17 May 2013 (UTC)

Correlation to active contributors and total blocks by month[edit]

Admittedly, this is not a complete dataset (based on active wikinews contributors per month) but a for the past couple of years, the data says there is no meaningful correlation between total administrative blocking actions and participation rates. --LauraHale (talk) 03:15, 6 May 2013 (UTC)

Month Year Blocks Active wikireporters
March-13 100 15
February-13 52 17
January-13 64 22
December-12 57 19
November-12 66 16
October-12 102 14
September-12 169 19
August-12 49 23
July-12 51 17
June-12 62 18
May-12 79 23
April-12 79 29
March-12 60 27
February-12 68 20
January-12 78 26
December-11 12 14
November-11 24 12
October-11 23 24
September-11 25 22
August-11 67 45
July-11 78 28
June-11 68 38
May-11 97 46
April-11 60 48
March-11 107 46
February-11 63 32
January-11 38 29
December-10 70 38
November-10 55 35
October-10 105 41
September-10 73 41
August-10 88 56
July-10 48 61
June-10 102 62
May-10 76 59
April-10 51 46
March-10 84 39
February-10 34 45
January-10 48 58
December-09 55 64
November-09 54 53
October-09 191 20
September-09 139 48
August-09 87 48
July-09 39 52
June-09 45 59
May-09 166 40
April-09 97 52
March-09 43 41
February-09 54 22
January-09 19 21
December-08 64 52
November-08 58 67
October-08 401 55
September-08 51 73
August-08 26 84
July-08 34 69
June-08 36 66
May-08 64 66
April-08 47 69
You've done ANOVA tests on this data, then? Ironholds (talk) 12:10, 17 May 2013 (UTC)